Lopiz 2014.

Study characteristics
Methods	Method of randomisation: used sequentially numbered, opaque, sealed envelopes Assessor blinding: not done Loss to follow‐up: 2 participants of the MultiLoc nail group were excluded (one died and one was lost to follow‐up)
Participants	Clinico San Carlos Hospital, Madrid, Spain Period of study recruitment: March 2011 to September 2012 54 participants with displaced Neer 2‐ or 3‐part proximal humeral fractures Exclusion criteria: pathological or open fractures, 4‐part fractures, concomitant fractures in the same upper limb, or the opposite and previous surgery on that shoulder. Lack of consent. Of 52: 41 female, 11 male; mean age 70 years, range 38 to 89 years
Interventions	All had general anaesthesia and intrascalene block. Mainly minimally invasive (percutaneous) ‐ small deltoid‐splitting incision (5 were open reduction with extended superior incision) 1. MultiLoc proximal humeral nail (MPHN) (Synthes‐DePuy, Solothurn, Switzerland) ‐ a straight nail 2. Polarus humeral nail (Acumed LLC, Hillsboro, OR, USA) ‐ a curved nail Postoperatively, participants were immobilised with a sling. Passive range‐of‐motion exercises were allowed 24 to 48 hours after surgery, followed as soon as possible by active‐assisted motion. Assigned: 28/26 Completed: 26/26 (mean 14 months)
Outcomes	Length of follow‐up: mean 14 months (6 to 22 months); formally 1, 3, 6 and 12 months Constant score (categories: excellent; good; satisfied; fair; poor) Constant score – adjusted for age and sex Physical tests to assess evidence of rotator cuff disease for entry point morbidity Non‐union, protrusion of the osteosynthesis material (subacromial impingement or articular surface intrusion of the screws), final alignment of the healed fracture (malunion) Reoperation (hardware removal for complications; reverse arthroplasty) Length of operation Intraoperative complications (none) Length of hospital stay Mortality
Funding and conflict of interest statements	Funding: not mentioned The authors declared that they and any research foundation with which they were affiliated received no financial payments or other benefits from any relevant commercial entity.
Notes	Request for information sent to Dr Lopiz on 20 October 2014 requesting clarification on method of sequence generation, details on the 2 excluded participants, query on tuberosity involvement of 3‐part fractures in the MultiLoc nail group, clarification on whether 1 of 2 participants had a reverse shoulder replacement and length of follow‐up times for each group. No response received.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"[A]ssistant generated the random allocation sequence, which was concealed from the authors." "Patients were randomly assigned to 2 parallel groups, initially at a 1:1 ratio"; description raises the concern that the sequence may have been predictable (not random) in the early stages ‐ but was probably OK
Allocation concealment (selection bias)	Low risk	"Randomization was carried out with use of sequentially numbered, opaque, sealed envelopes." "All patients were randomized by a research co‐ordinator who was not involved subsequently in the study."
Blinding (performance bias and detection bias) Functional outcomes, pain, clinical outcomes, complications	High risk	“The health care providers involved with subsequent patient care were not blinded to the treatment.” No mention of independent or blinded outcome assessment
Blinding (performance bias and detection bias) Death, reoperation	Unclear risk	“The health care providers involved with subsequent patient care were not blinded to the treatment.” However, it is unlikely that lack of blinding will affect the reporting of these outcomes.
Incomplete outcome data (attrition bias) Functional outcomes, pain, clinical outcomes, complications	Unclear risk	Variable follow‐up with no confirmation of similar follow‐ups in the two groups. Additionally, data lost from 2 participants in the Polarus nail group (1 died + 1 lost to follow‐up).
Incomplete outcome data (attrition bias) Death, reoperation	Low risk	Data lost from 2 participants in the Polarus nail group (1 died + 1 lost to follow‐up). Unlikely to bias the results
Selective reporting (reporting bias)	Unclear risk	Insufficient information to judge this. No protocol found
Balance in baseline characteristics?	Low risk	Although baseline data were not presented for 2 participants in the MultiLoc nailing group, there were no major imbalances in baseline characteristics between the two groups: "No statistically significant differences were found between the 2 groups."
Free from performance bias?	Low risk	"All surgeries were performed by 1 of the 3 senior trauma surgeons in the unit." Postoperative care was the same in both groups.